Hello,
After performing Statistical Analysis in my dataset and downloading the related files, there is a file named “data_processed”. What exactly is this file and where is it used in the Analysis section ?
I understand that while it doesn’t contain the original dataset it should at least maintain the correct pattern of the metabolites in the dataset. However, I calculated the ratios of the same metabolite across different samples and compared them to the original dataset, and while the majority are correct, some are wrong. I provide the examples below for 2 distinct metabolites:
Example 1 (metabolite 1):
Ratios from the original dataset: 3.7 13.8 10.5
Ratios from the processed dataset: 3.73 0.138 1.05
Example 2 (metabolite 2):
Ratios from the original dataset: 1.2 1.4 1.3
Ratios from the processed dataset: 1.2 14.0 1.3
In example 1 the first ratio is correct, but the other two deviate by a factor of /100.
In example 2 the first and third ratios are correct, but the second deviates by a factor of x10.
If normalization and further statistical analysis are perfomed on the “processed” file they will mistakenly capture the metabolite patterns. Specifically, i noticed this in the heatmap view, when using both options for the data source (normalized and original) for these specific metabolites that i mentioned in the above example.
Is there something I understand wrong, or otherwise, what would be the reason of wrong pattern in the heatmap even when using the original data? I know that a specific metabolite should appear as upregulated in a specific condition based on my original data, but in the heatmap view (using original data) it appears the opposite.
I would appreciate it a lot if you can help me figure this out.
EDIT; In the test file provided by MetaboAnalyst “Metabolite concentrations of 77 urine samples from cancer patients measured by 1H NMR (Eisner R, et al.). Group 1- cachexic; group 2 - control” no change was observed between the original and processed files, in contrast to my dataset.